ABSTRACT
The relationships between water chemistry and watershed physical characteristics
were examined for twelve high elevation lakes in Grand Teton National Park (GRTE),
Wyoming. Twelve lakes were sampled for major anions, cations, ANC, alkalinity,
conductance, pH, DOC, total P, and total N in order to determine their sensitivity to
acidification, and to understand the relationship between basin characteristics and lake
water chemistry. A step-wise multiple linear regression method was used in an attempt to
associate basin characteristics with water chemistry and to identify those independent
variables that exert a strong influence on the association. Discriminant analysis was also
used to identify which variables make a significant contribution to the classification of lake
sensitivity. The results indicate that many of the high elevation lakes in GRTE are sensitive
to acidification, with half of the lakes having ANC concentrations < 100µeq L-1. Lakes in
basins with granitic and/or metamorphic bedrock are the most sensitive to acidification,
particularly when the basin contains a high proportion of young debris. Lakes with basins
that are at least partially underlain by limestone bedrock are the least sensitive to
acidification, regardless of the presence of young debris. Granite and limestone served as
the best predictors for solute concentrations with young debris and steep slopes playing
significant roles for most solutes – especially major base cations, pH, and acidity. The
regression models for alkalinity and ANC were excellent predictors for concentrations of
both variables and explained 86.5% of the variance in ANC and 94.8% of the variance in
alkalinity. Glacier dissolution in GRTE study basins may be responsible for seasonal
increases in NO3- concentrations in glacier-fed lakes, which in turn decreases the ANC.
Additional monitoring of GRTE high elevation lakes is required in order to provide more
information on long-term water quality status and trends.
GLACIAL MELTWATER
Atmospheric Deposition Effects on High Alpine
Lakes in Grand Teton National Park, Wyoming
Jennifer Corbin and Scott Woods
University of Montana, Department of Ecosystem and Conservation Sciences
RESULTS – SOLUTE CONCENTRATIONS
14
N=5
Variability in solute concentrations
10
METHODS
Sample Collection
Seasonal Trends in Solute Concentrations
NO3 (ueq/L)
6
Glacier-fed
4
2
0
A
M
J
J
Delta Lake – ANC 42.5 µeq L-1
GRTE Mechanisms = ANC of limestone AND ↑ nitrate in talus fields
Carbonate Rock Dissolution
Limestone is present in parts of the study area, and lakes with limestone bedrock appear to have sufficient buffering
capacity as a result of carbonate weathering. The three basins with limestone bedrock, Snowdrift Lake (Figure 2a and
2b), Sunset Lake, and Alaska Basin Lake, had ANC values of 676.2, 1488.3 and 110.3µeq L-1, respectively, for a mean of
758.3 µeq L-1. In contrast, lakes without limestone bedrock had ANC values ranging from 42.5 to 219.6 µeq L-1 with a
mean of 89.3 µeq L-1.
Temporal Trends in Solute Concentrations
Nitrate in Talus Fields
•No consistent ANC trends
•7 increase (Bradley Lake - 102.1 to 148.9 µeq L-1)
•5 decrease (Mica Lake - 149.4 to 77.9 µeq L-1)
•Nitrate increase
The results of the present study suggest that both mechanisms – the acid neutralizing effect of limestone bedrock, and high nitrate from
talus fields – affect the basin water chemistry at sites in GRTE. The three basins with limestone bedrock included in the present study
(Alaska Basin, Snowdrift and Sunset Lakes) also have high proportions of young debris, and ANC values are relatively high,
suggesting that a similar mechanism to that proposed by Gulley and Parker (1985) for Schoolroom Lake is also controlling the ANC in
these lakes. However, watersheds without limestone but with a large amount of young debris, such as Lake Solitude and Mica Lake,
have some of the lowest ANC values. Snowpack studies have shown that the neutralizing effect of Ca2+ is sometimes overcome by
increases in NO3- and SO42- (Mast et al 2001), which is illustrated by lower pH values (Turk et al 2001). The results of the present
study indicate that, in watersheds without limestone, high NO3 increases the sensitivity to acidification.
Figure 1 – Acid Neutralizing Capacity (ANC) Trends
in Trapper Lake (1985 - Landers et al., 1996 - Williams
and Tonnessen, 2002 - GRTE Survey)
500
Trends since 1985:
• Only
one lake (Trapper Lake) has data for 1985, 1995 and 2002
•ANC has decreased from 441 to 206 me/L (Figure 1)
• Major cations have decreased
• NO3 has increased
• SO4 has increased
400
300
200
1980
RESULTS – PREDICTION SUCCESS
1990
2000
2010
800
CONCLUSION
100
Predicted Mg
600
Figure2b - Limestone cliffs above Snowdrift Lake
O
7
Figure 3 - Relations between glacier-fed lakes
and seasonal mean NO3- concentrations.
YEAR
80
400
60
200
40
0
0
200
400
600
800
Percent Granite
Measured Mg
2000
Predicted ANC
Figure 2a - Snowdrift Lake – ANC 1488 µeq L-1
S
N
N=7
=
TALUS EFFECTS
On average:
•Ca2+ and Mg2+ concentrations decreased
•Na+ concentrations increased
•Phosphate concentrations were below detection limits
•Nitrate concentrations decreased seasonally
•SO42- concentrations decreased seasonally
Modeling efforts for the 2002 study in GRTE were centered primarily
on multiple linear regression analysis and discriminant analysis. Results of
the multiple linear regression models for study lakes were variable, but, on
average, showed good agreement with earlier sample data (Figure 1a and
2b). Granite and limestone served as the best predictors for solute
concentrations with young debris and steep slopes playing significant roles
for most solutes – especially major base cations, pH, and acidity.
In the discriminant analysis, the variable that best defined group
membership was granite (Figure 2). Typically, lakes with ANC
concentrations < 100µeq L-1 resided in basins that had total granite
compositions ranging from 20% to 80%. When the categories were
applied to basins sampled in 1996, the analysis held true to the
categorization. South Leigh Lake had an ANC concentration of 83.2 µeq
L-1 and was underlain with granite deposits totaling approximately 72%
of the entire basin. Likewise, Lake of the Crags had an ANC of 67 µeq L-1
and resided in a basin of 22% granite. Therefore, the results of the
discriminate analysis support the regression analysis results in that granite
appears to be a suitable predictor for ANC concentrations in GRTE.
A
MONTH
ANC (ueq/L)
•Basins were categorized based on topography, geology, and habitat and vegetation cover types.
•Categories were expressed as a percentage of total basin area.
•A step-wise multiple linear regression method was used in the GRTE study in an attempt to
associate basin characteristics with water chemistry and to identify those independent variables that
exert a strong influence on the association.
•Discriminant analysis was also used to identify which variables make a significant contribution to
the classification of lake sensitivity.
•The model was calibrated with the data collected in the summer of 2002 at GRTE. Mean
concentrations of late season samples were entered into the model.
•Water chemistry data collected by Williams and Tonnessen in 1996 and by Landers et al. in 1985
were used for model testing and validation.
Not glacier-fed
8
Trends since 1995:
Model Development
12
Teton Glacier
•Wide range of ANC
•Range = 37.9 to 1488.3 µeq L-1
•Median = 92.2 µeq L-1
•Base cations
•Median = 76.8µeq L-1 for Ca2+
•Median = 28.9µeq L-1 for Mg2+
•Conductivity – median = 12.1µS cm-1
•NO3•Range = 0.1 µeq L-1 to 20.1 µeq L-1
•Median = 8.8 µeq L-1
•DOC
•Range = 289 ppb to 1776 ppb
•Median = 712 ppb
•Lake selection was based on past sampling and morphometric characteristics
•Twelve lakes selected:
•2100m – 3000m elevation
•0.9ha – 27ha surface area
•Early- and late-season samples were taken at each lake.
•Analytes: ANC – Acid Neutralizing Capacity, Alkalinity, [pH], Specific Conductance, Major
Cations, Major Anions, DOC – Dissolved Organic Carbon, TP – Total (and particulate)
Phosphorous, TN – Total (and particulate) Nitrogen
Seasonal melt from remnant glaciers feeding several of the lakes in GRTE may increase
sensitivity to acidification by increasing the nitrogen flux in the late summer (Figure 3). Delta
Lake – a glacier-fed lake – had a mean Ca2+ concentration of 50.9µeq L-1 but NO3- and
SO42- concentrations were high (20.1 µeq L-1 and 12.3 µeq L-1, respectively), resulting in an
ANC value of 42.5 µeq L-1. In contrast, Alaska Basin Lake had a mean Ca2+ concentration of
68.5µeq L-1, a mean NO3- concentration of 0.4µeq L-1, and a mean SO42- concentration of
13.7µeq L-1. The ANC value for this lake was 110.3 µeq L-1.
1500
20
0
-20
N=
1000
3
3
6
< 50ue/L
< 100ue/L
>100ue/L
Sensitivity
500
0
0
500
1000
1500
2000
Measured ANC
Figure 1 – Measured and predicted values of (a)
Mg2+ and (b) ANC
Figure 2 - Boxplot of percent granite in study
basins relative to acidification susceptibility.
•Sampled 12 lakes
•GRTE lakes are sensitive
•7 lakes < 100µeq L-1
•Granitic bedrock = sensitive
•Granitic bedrock + young debris = most sensitive
•Limestone bedrock = least sensitive
•Glacial meltwater = increased sensitivity
•Trends indicate a consistent decline in ANC
•Recommend additional monitoring